Pith. sign in

REVIEW 4 major objections 4 minor 38 references

Cascading Bandits Robust to Adversarial Corruptions

T0 review · 4 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Online learning to rank in the cascade model survives adversarial click corruption: the paper proves two algorithms keep logarithmic regret without attacks and lose only a linear-in-C factor under a bounded adversary.

desk verdict A sensible new problem and plausible algorithms, but Lemma 1's independence assumption lets the adversary concentrate all corruption on the slow instance, so the main claim needs a major fix. read the letter →

arxiv 2502.08077 v1 pith:JWBOXQPS submitted 2025-02-12 cs.LG

Jize Xie , Cheng Chen , Zhiyong Wang , Shuai Li This is my paper · ORCID
classification cs.LG
keywords cascadingbanditsadversarialcorruptionslearningtorankclickfraudregretboundseliminationalgorithmsonlinerobustness
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Online learning to rank in the cascade model—where a user scans a short list and clicks the first attractive item—breaks down if part of the click feedback is adversarially corrupted. This paper claims the damage can be contained: it introduces CascadeRKC and CascadeRAC, for known and unknown corruption budgets C, and proves both keep the clean-environment logarithmic regret while paying only an extra term linear in C under an adaptive adversary. The proofs rest on a position-based elimination rule that keeps a separate eliminated set for each list position, plus a two-instance design in which a slow, corruption-tolerant instance is played rarely so it absorbs most of the attack. If the bounds are right, click fraud in recommender systems no longer forces a choice between logarithmic no-attack regret and robustness.

What carries the argument

Position-based elimination (PBE) maintains one eliminated set $M^k$ for each of the $K$ list positions instead of one global set, removing item $e$ from position $k$ as soon as the lower confidence bound of $k$ better items beats the upper confidence bound of $e$. The robust wrapper runs two PBE instances: a fast instance $F$ with tight confidence radius $wd_F(a)=\sqrt{\log(8LT/\delta)/T^F(a)}$ and a slow instance $S$ with enlarged radius $wd_S(a)=\sqrt{\log(8LT/\delta)/T^S(a)} + 2\log(8LT/\delta)/T^S(a)$. Instance $S$ is played with probability $1/C$, making the expected corruption on $S$ a constant, and every elimination in $S$ is copied to $F$. This carries the argument because the enlarged radius lets $S$ tolerate up to $O(\log T)$ corrupted observations while the $1/C$ sampling keeps suboptimal items from being played often enough to inflate regret.

What would settle it

Build an adversary that observes the recommended list each round and corrupts feedback only on rounds where the slow instance is active, then measure CascadeRKC's cumulative regret as C grows; if the slow instance's observed corruption exceeds O(log T) with high probability or if regret grows faster than O(C $log^{2}$ T), the independence premise behind Lemma 1 is false for adaptive adversaries that condition on the active instance.

Watch

Extended reading notes

Core claim

The paper's central claim is that a total corruption budget C in cascading bandits can be absorbed with only a linear price in C. Concretely, CascadeRKC achieves, with high probability, $R(T) \le O(\sum_{e=K+1}^{L} (1/\Delta_{e,K}) \, K L C \, (\log(LT/\delta))^2)$, and when the corruption level is unknown, CascadeRAC achieves $R(T) \le O(\sum_{e=K+1}^{L} K (L C \log(LT/\delta) + \log T) \log(LT/\delta) / \Delta_{e,K})$. With no corruption both reduce to gap-dependent logarithmic regret. The mechanism is elimination-based: the slow instance is sampled with probability $1/C$, so only a constant amount of corruption reaches it in expectation, and the fast instance inherits its eliminations to stay efficient. When the list size is $K=1$, the bounds coincide with known robust multi-armed bandit guarantees, which the paper reads as tightness of the corruption term.

Load-bearing premise

The proof that the slow instance sees only O(1) corruption assumes the adversary's corruptions hit the fast and slow instances in proportion to how often each is played; the adversary is handed the recommended list, which can reveal which instance is active, so a determined attacker could funnel all C corruptions into the slow instance and break the logarithmic-regret argument.

Editorial extensions

If this is right

  • If Theorem 3 holds, a platform that knows its corruption budget $C$ can run CascadeRKC and keep logarithmic regret when no attack is happening, degrading only linearly in $C$.
  • If Theorem 4 holds, no prior knowledge of $C$ is needed: CascadeRAC stays within $O(C \log^2 T / \Delta)$ regret, with an extra $\log T$ factor for searching over $\log T$ instance levels.
  • Both algorithms eliminate items per position rather than globally, which is what keeps the clean-environment regret logarithmic instead of $\sqrt{T}$.
  • At $K=1$ the bounds match the known robust stochastic bandit bounds, indicating the corruption term is tight up to logarithmic factors.
  • The experimental comparisons show the two algorithms outperform standard cascading-bandit baselines under periodic, early-phase, and varying-gap corruption on synthetic and real-world data.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A determined adversary could exploit the fact that the recommended list reveals which instance is playing, concentrating all $C$ corruptions on the slow instance and violating the $O(1)$ corruption premise of Lemma 1; randomizing between two behaviorally indistinguishable lists would close that loophole.
  • The two-instance PBE construction should transfer to position-based click models or other partial-feedback ranking settings, since PBE only requires per-position confidence intervals.
  • A concrete testable extension is to run CascadeRKC against an adversary that always corrupts rounds when the slow instance is active and measure whether regret becomes superlinear in $C$; this would show whether the independence assumption in Lemma 1 is actually needed.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper introduces the Cascading Bandits with Adversarial Corruptions (CBAC) problem, in which an adaptive adversary with total corruption budget C can modify click feedback. It proposes two algorithms, CascadeRKC (for known C) and CascadeRAC (for agnostic C), built on a position-based elimination (PBE) subroutine. The paper claims gap-dependent logarithmic regret in corruption-free settings and regret growing linearly in C under corruption, with proofs deferred to an appendix and experiments on synthetic and real-world datasets.

Significance. If the theorems were correct, this would be the first treatment of adversarial corruptions in cascading bandits, and the position-based elimination rule combined with multi-instance sampling is a natural extension of Lykouris et al. The claimed O(C) dependence on the corruption level is also consistent with the known MAB lower bound. However, the central adaptive-corruption guarantee rests on an independence assumption in Lemma 1 that conflicts with the paper's own protocol; the appendix proofs are sketches with typos and unstated steps. The experiments, while suggestive, use only fixed-schedule adversaries and therefore do not exercise the claimed adaptive robustness.

major comments (4)
  1. [Section 5.1 / Appendix 9] Lemma 1 is not valid under the protocol of Section 3. The proof writes E[G_t(a_{S,t})|H_{t-1}] = C_{a_{S,t}}/C, which presumes that the adversary's corruption decision is independent of whether instance S was sampled in round t. But the protocol explicitly gives the attacker the recommended list A_t before the corruption is chosen, and A_t reveals the sampled instance because the two instances maintain different elimination sets (F eliminates items faster, so M^F is a superset of M^S). An adversary can therefore corrupt only on rounds in which S was active, concentrating all C corruptions on S and making its corruption as large as C rather than O(log(1/delta)). Since Lemma 2 and Theorems 3-4 inherit this bound, the main theoretical claims are unsupported for the stated adaptive adversary.
  2. [Appendix 10] The proof of Lemma 2 confuses delta1 and delta2 and misstates the Hoeffding radius. After selecting delta2 = delta'/LT, the union bound should give a radius involving log(2LT/delta')/T^S(a), not log(2/delta')/T^S(a); the subsequent step then mixes Cs <= log(1/delta1)+3 with delta2 without a coherent high-probability statement. These errors make Eq. (3) and the claim that optimal items are never eliminated insufficiently specified.
  3. [Section 4.2 / Algorithm 2] Algorithm 2 samples the slow instance S with probability 1/C, which is undefined when C = 0. Since the paper also claims logarithmic regret in the corruption-free case, the algorithm must handle C = 0 explicitly, for example by running only the fast instance F; as written, the pseudo-code and the subsequent analysis do not cover the no-corruption regime.
  4. [Appendix 11 / Proof of Theorem 3] The bound on the F-instance regret is not derived. The proof asserts that a sub-optimal item e appears at position k in F at most L*C*T^S(e) times because 'every move in the slow active arm elimination occurs with probability 1/C and, at least 1/L of these moves are plays of e,' but no justification is given for the 1/L factor in the cascade model, where elimination is position-specific and the two instances use different active sets. This step is load-bearing because the factor L appears in Theorem 3's bound.
minor comments (4)
  1. [Section 2] The section title contains a typo: 'Relatet Works' should be 'Related Works'.
  2. [Section 6 / Figures] The axis labels and legends in Figures 1-5 appear as unicode artifacts, making the plots difficult to read; the figures should be regenerated with standard text fonts.
  3. [Table 1] In the Movielens column, the entry '18,6502' for CascadeKL-UCB is likely a typo and should probably read '186,502'.
  4. [Section 5.1 / Lemma 1] The phrase 'during its exploration phase' is not defined either in the lemma statement or in the proof, which actually bounds corruption over all rounds t; please clarify the intended event.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: regret bounds are derived from the algorithms' definitions and external concentration lemmas, not from fitted parameters or self-citations.

full rationale

The proof chain is self-contained with respect to the paper's own claims. Theorems 3 and 4 bound regret using the elimination rules of the proposed PBE instances, the confidence radii, and standard external tools (Hoeffding inequalities, Theorem 1 of Kveton et al. 2015a, and Lemma B.1 of Lykouris et al. 2018). No fitted parameter is renamed as a prediction, and no equation is defined in terms of the result it purports to prove. The algorithms are explicitly constructed to control sub-optimal plays, and the resulting bounds are derived rather than assumed. Citations to the authors' prior works (e.g., Wang et al. 2024, Dai et al. 2024, Liu et al. 2021) appear only as related work and are not load-bearing for the main theorems. The proof of Lemma 1 does assume that the adversary's corruption decision is independent of whether the S or F instance is sampled, which is not justified under the stated adaptive-adversary protocol and is a genuine correctness risk; however, that is a soundness gap, not a circular reduction, because the claimed bound is not equivalent to an input by construction. The degeneration check to K=1 matches, rather than repackages, the known MAB-with-corruptions bounds.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The theory introduces no fitted constants or invented entities. The main burden lies in domain assumptions about the cascade model and the corruption process, plus the questionable independence assumption in Lemma 1. The concentration lemmas and regret decomposition are standard external tools.

assumptions (5)
  • domain assumption Assumption 1: attraction indicators are independent Bernoulli random variables with means w(a).
    Stated in Section 3; used to express expected reward as f(A, w_A) and to enable the regret decomposition in the proofs.
  • standard math Standard concentration inequalities, including Hoeffding's inequality and the martingale bound Lemma B.1 of Lykouris et al. 2018.
    Used throughout the appendix proofs of Lemmas 1 and 2 to control estimation errors and corruption deviations.
  • domain assumption The adversary's per-round corruption is binary: the maximum over positions of |R_t(a_k) - Rtilde_t(a_k)| is either 0 or 1.
    The corruption budget in Section 3 sums these maxima, and the proof of Lemma 1 assumes each corrupted round contributes exactly one unit, so C_{a,t} ∈ {0,1}.
  • ad hoc to paper Expected corruption in the slow instance S is O(1) because S is sampled with probability 1/C, which presumes the adversary's corruption decision is independent of whether S or F is selected.
    Appendix Section 9; this is the load-bearing premise not justified by the stated protocol, since the adversary observes the recommended list At and may infer the active instance.
  • standard math Regret decomposition via Theorem 1 of Kveton et al. 2015a, bounding immediate regret by a sum over suboptimal items and positions.
    Invoked in the proof of Theorem 3 to translate per-item elimination events into regret terms.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Cascading Bandits Robust to Adversarial Corruptions." pith.science (2026). https://pith.science/paper/JWBOXQPS

@misc{pith2026250208077,
  author       = {Pith},
  title        = {Pith review of: Cascading Bandits Robust to Adversarial Corruptions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JWBOXQPS}},
  note         = {Machine review of arXiv:2502.08077}
}
read the original abstract

Online learning to rank sequentially recommends a small list of items to users from a large candidate set and receives the users' click feedback. In many real-world scenarios, users browse the recommended list in order and click the first attractive item without checking the rest. Such behaviors are usually formulated as the cascade model. Many recent works study algorithms for cascading bandits, an online learning to rank framework in the cascade model. However, the performance of existing methods may drop significantly if part of the user feedback is adversarially corrupted (e.g., click fraud). In this work, we study how to resist adversarial corruptions in cascading bandits. We first formulate the ``\textit{Cascading Bandits with Adversarial Corruptions}" (CBAC) problem, which assumes that there is an adaptive adversary that may manipulate the user feedback. Then we propose two robust algorithms for this problem, which assume the corruption level is known and agnostic, respectively. We show that both algorithms can achieve logarithmic regret when the algorithm is not under attack, and the regret increases linearly with the corruption level. The experimental results also verify the robustness of our methods.

Figures

Figures reproduced from arXiv: 2502.08077 by the authors.

Figure 1
Figure 1. Comparison of cumulative regrets on the synthetic dataset with [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Cumulative regret on the real datasets. (a) shows the results in the Yelp dataset, (b) shows the results in the [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Cumulative regret under different ∆ between optimal items and the target item. (a) exhibits the results when ∆ = 0.1, (b) exhibits the results when ∆ = 0.2, and (c) exhibits the results when ∆ = 0.4. Radlinski et al. [2008]. Here, the first three baselines are all UCB-based cascading bandits algorithms, and they only differ in the choices of the computation of UCB. They are widely used and perform well in the stocha… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Cumulative regret under different corruption levels in the synthetic dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Cumulative regret under different corruption levels in the Movielens dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 35 canonical work pages

  1. [1]

    Cascading bandits: Learning to rank in the cascade model

    Branislav Kveton, Csaba Szepesvari, Zheng Wen, and Azin Ashkan. Cascading bandits: Learning to rank in the cascade model. In International conference on machine learning, pages 767--776. PMLR, 2015 a

    work page 2015

  2. [2]

    Learning to rank: Regret lower bounds and efficient algorithms

    Richard Combes, Stefan Magureanu, Alexandre Proutiere, and Cyrille Laroche. Learning to rank: Regret lower bounds and efficient algorithms. In Proceedings of the 2015 ACM SIGMETRICS international conference on measurement and modeling of computer systems, pages 231--244, 2015

    work page 2015

  3. [3]

    Learning to rank: from pairwise approach to listwise approach

    Zhe Cao, Tao Qin, Tie-Yan Liu, Ming-Feng Tsai, and Hang Li. Learning to rank: from pairwise approach to listwise approach. In Proceedings of the 24th international conference on Machine learning, pages 129--136, 2007

    2007

  4. [4]

    Learning to rank

    Andrew Trotman. Learning to rank. Information Retrieval, 8: 0 359--381, 2005

    work page 2005

  5. [5]

    Toprank: A practical algorithm for online stochastic ranking

    Tor Lattimore, Branislav Kveton, Shuai Li, and Csaba Szepesvari. Toprank: A practical algorithm for online stochastic ranking. Advances in Neural Information Processing Systems, 31, 2018

    work page 2018

  6. [6]

    Thompson sampling algorithms for cascading bandits

    Zixin Zhong, Wang Chi Chueng, and Vincent YF Tan. Thompson sampling algorithms for cascading bandits. The Journal of Machine Learning Research, 22 0 (1): 0 9915--9980, 2021

    work page 2021

  7. [7]

    An experimental comparison of click position-bias models

    Nick Craswell, Onno Zoeter, Michael Taylor, and Bill Ramsey. An experimental comparison of click position-bias models. In Proceedings of the 2008 international conference on web search and data mining, pages 87--94, 2008

    work page 2008

  8. [8]

    Combinatorial cascading bandits

    Branislav Kveton, Zheng Wen, Azin Ashkan, and Csaba Szepesvari. Combinatorial cascading bandits. Advances in Neural Information Processing Systems, 28, 2015 b

    work page 2015

Show all 38 references
  1. [9]

    Cascading non-stationary bandits: Online learning to rank in the non-stationary cascade model

    Chang Li and Maarten De Rijke. Cascading non-stationary bandits: Online learning to rank in the non-stationary cascade model. arXiv preprint arXiv:1905.12370, 2019

    1905 arXiv

  2. [10]

    Stochastic bandits robust to adversarial corruptions

    Thodoris Lykouris, Vahab Mirrokni, and Renato Paes Leme. Stochastic bandits robust to adversarial corruptions. In Proceedings of the 50th Annual ACM SIGACT Symposium on Theory of Computing, pages 114--122, 2018

    2018

  3. [11]

    Better algorithms for stochastic bandits with adversarial corruptions

    Anupam Gupta, Tomer Koren, and Kunal Talwar. Better algorithms for stochastic bandits with adversarial corruptions. In Conference on Learning Theory, pages 1562--1578. PMLR, 2019

    2019

  4. [12]

    Learning diverse rankings with multi-armed bandits

    Filip Radlinski, Robert Kleinberg, and Thorsten Joachims. Learning diverse rankings with multi-armed bandits. In Proceedings of the 25th international conference on Machine learning, pages 784--791, 2008

    2008

  5. [13]

    Cooperative stochastic multi-agent multi-armed bandits robust to adversarial corruptions

    Junyan Liu, Shuai Li, and Dapeng Li. Cooperative stochastic multi-agent multi-armed bandits robust to adversarial corruptions. arXiv preprint arXiv:2106.04207, 2021

    2021 arXiv

  6. [14]

    Stochastic graphical bandits with adversarial corruptions

    Shiyin Lu, Guanghui Wang, and Lijun Zhang. Stochastic graphical bandits with adversarial corruptions. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 8749--8757, 2021

    2021

  7. [15]

    Robust stochastic linear contextual bandits under adversarial attacks

    Qin Ding, Cho-Jui Hsieh, and James Sharpnack. Robust stochastic linear contextual bandits under adversarial attacks. In International Conference on Artificial Intelligence and Statistics, pages 7111--7123. PMLR, 2022

    2022

  8. [16]

    Online corrupted user detection and regret minimization

    Zhiyong Wang, Jize Xie, Tong Yu, Shuai Li, and John Lui. Online corrupted user detection and regret minimization. Advances in Neural Information Processing Systems, 36, 2024

    2024

  9. [17]

    Action elimination and stopping conditions for the multi-armed bandit and reinforcement learning problems

    Eyal Even-Dar, Shie Mannor, Yishay Mansour, and Sridhar Mahadevan. Action elimination and stopping conditions for the multi-armed bandit and reinforcement learning problems. Journal of machine learning research, 7 0 (6), 2006

    2006

  10. [18]

    Online learning to rank in stochastic click models

    Masrour Zoghi, Tomas Tunys, Mohammad Ghavamzadeh, Branislav Kveton, Csaba Szepesvari, and Zheng Wen. Online learning to rank in stochastic click models. In International conference on machine learning, pages 4199--4208. PMLR, 2017

    2017

  11. [19]

    Minimax regret for cascading bandits

    Daniel Vial, Sujay Sanghavi, Sanjay Shakkottai, and R Srikant. Minimax regret for cascading bandits. arXiv preprint arXiv:2203.12577, 2022

    2022 arXiv

  12. [20]

    Contextual combinatorial cascading bandits

    Shuai Li, Baoxiang Wang, Shengyu Zhang, and Wei Chen. Contextual combinatorial cascading bandits. In International conference on machine learning, pages 1245--1253. PMLR, 2016

    2016

  13. [21]

    Online clustering of contextual cascading bandits

    Shuai Li and Shengyu Zhang. Online clustering of contextual cascading bandits. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32, 2018

    2018

  14. [22]

    Cascading bandits for large-scale recommendation problems

    Shi Zong, Hao Ni, Kenny Sung, Nan Rosemary Ke, Zheng Wen, and Branislav Kveton. Cascading bandits for large-scale recommendation problems. arXiv preprint arXiv:1603.05359, 2016

    2016 arXiv

  15. [23]

    Online learning to rank with features

    Shuai Li, Tor Lattimore, and Csaba Szepesv \'a ri. Online learning to rank with features. In International Conference on Machine Learning, pages 3856--3865. PMLR, 2019 a

    2019

  16. [24]

    Cascading hybrid bandits: Online learning to rank for relevance and diversity

    Chang Li, Haoyun Feng, and Maarten de Rijke. Cascading hybrid bandits: Online learning to rank for relevance and diversity. In Proceedings of the 14th ACM Conference on Recommender Systems, pages 33--42, 2020

    2020

  17. [25]

    Corruption-tolerant bandit learning

    Sayash Kapoor, Kumar Kshitij Patel, and Purushottam Kar. Corruption-tolerant bandit learning. Machine Learning, 108 0 (4): 0 687--715, 2019

    2019

  18. [26]

    Tsallis-inf: An optimal algorithm for stochastic and adversarial bandits

    Julian Zimmert and Yevgeny Seldin. Tsallis-inf: An optimal algorithm for stochastic and adversarial bandits. The Journal of Machine Learning Research, 22 0 (1): 0 1310--1358, 2021

    2021

  19. [27]

    Stochastic dueling bandits with adversarial corruption

    Arpit Agarwal, Shivani Agarwal, and Prathamesh Patil. Stochastic dueling bandits with adversarial corruption. In Algorithmic Learning Theory, pages 217--248. PMLR, 2021

    2021

  20. [28]

    Adversarial bandits with corruptions: Regret lower bound and no-regret algorithm

    Mohammad Hajiesmaili, Mohammad Sadegh Talebi, John Lui, Wing Shing Wong, et al. Adversarial bandits with corruptions: Regret lower bound and no-regret algorithm. Advances in Neural Information Processing Systems, 33: 0 19943--19952, 2020

    2020

  21. [29]

    Stochastic linear optimization with adversarial corruption

    Yingkai Li, Edmund Y Lou, and Liren Shan. Stochastic linear optimization with adversarial corruption. arXiv preprint arXiv:1909.02109, 2019 b

    1909 arXiv

  22. [30]

    Nearly optimal algorithms for linear contextual bandits with adversarial corruptions

    Jiafan He, Dongruo Zhou, Tong Zhang, and Quanquan Gu. Nearly optimal algorithms for linear contextual bandits with adversarial corruptions. arXiv preprint arXiv:2205.06811, 2022

    2022 arXiv

  23. [31]

    Online learning and detecting corrupted users for conversational recommendation systems

    Xiangxiang Dai, Zhiyong Wang, Jize Xie, Tong Yu, and John CS Lui. Online learning and detecting corrupted users for conversational recommendation systems. IEEE Transactions on Knowledge and Data Engineering, 2024

    2024

  24. [32]

    Stochastic linear bandits robust to adversarial attacks

    Ilija Bogunovic, Arpan Losalka, Andreas Krause, and Jonathan Scarlett. Stochastic linear bandits robust to adversarial attacks. In International Conference on Artificial Intelligence and Statistics, pages 991--999. PMLR, 2021

    2021

  25. [33]

    The movielens datasets: History and context

    F Maxwell Harper and Joseph A Konstan. The movielens datasets: History and context. Acm transactions on interactive intelligent systems (tiis), 5 0 (4): 0 1--19, 2015

    2015

  26. [34]

    Federated multi-armed bandits

    Chengshuai Shi and Cong Shen. Federated multi-armed bandits. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 9603--9611, 2021

    2021

  27. [35]

    Comparison-based conversational recommender system with relative bandit feedback

    Zhihui Xie, Tong Yu, Canzhe Zhao, and Shuai Li. Comparison-based conversational recommender system with relative bandit feedback. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 1400--1409, 2021

    2021

  28. [36]

    Bernoulli rank- 1 bandits for click feedback

    Sumeet Katariya, Branislav Kveton, Csaba Szepesv \'a ri, Claire Vernade, and Zheng Wen. Bernoulli rank- 1 bandits for click feedback. arXiv preprint arXiv:1703.06513, 2017

    2017 arXiv

  29. [37]

    Clustering of conversational bandits for user preference learning and elicitation

    Junda Wu, Canzhe Zhao, Tong Yu, Jingyang Li, and Shuai Li. Clustering of conversational bandits for user preference learning and elicitation. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management, pages 2129--2139, 2021

    2021

  30. [38]

    Click models for web search

    Aleksandr Chuklin, Ilya Markov, and Maarten de Rijke. Click models for web search. Synthesis lectures on information concepts, retrieval, and services, 7 0 (3): 0 1--115, 2015

    2015

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.